This sort of stuff is why I quit Intel and sold all my stock. In my opinion (and I haven’t been following post-Kersanich Intel closely) Intel should be focusing on their process and processors, I don’t think the execs really have the will it would take to diversify.
These acquisitions almost always end up being run into the ground because Intel is so used to turning the crank on their mature (read: awesomely ubiquitous software support) CPU architecture that they don’t want to spend much on R&D on things that aren’t proven money makers. But they keep trying to div diversify in a very half-assed way because they have such a high market share in CPUs (not that it is without legitimate threat from AMD) that to continue to grow they want to hit new markets.
I actually disagree with your analysis, my experience with Intel was from an acquisition. They decided they were going to do 'big bets' which involved massively massively investing in a few areas. What happened, was they acquired the company and told us to grow in certain areas enormously and quickly. So it's not that they acquired and then cut R&D. The opposite happened, they hugely increased R&D (sure they cut lots of other things that caused trouble), but it was like winning the lottery, suddenly everyone could hire for everything. So the message was "Get on and hire, because it won't be like this forever, and if you fail your hiring target, you'll be stuck delivering big projects with no staff". Immediately hiring became a mess, the weight of bringing on all this new staff massively slowed everything down. In most ways we ended up delivering less with twice the staff. Needless to say: this didn't produce the results they wanted.
Fortunately for my team, Intel completely shitting the bed on 10nm meant that all these problems were barely noticed - who can blame you if the silicon isn't there.
The thing is, that when you pay a premium, overinvest and under-deliver you end up with no choice in the exec team - and you end up 'cutting you losses' and doing a McAfee.
It's amazing to me how often this story repeats itself in tech, yet the results are almost always the same:
1. Old, hugely successful company needs to find new growth areas.
2. Company acquires company in high growth area (sometimes innovative, sometimes not)
3. Taking from the MBA playbook that higher investment equals more growth, acquired company is given aggressive targets but also a huge budget to meet them.
4. Company goes on a massive, usually relatively untargeted hiring spree, things become a major shitshow, acquired leadership team is eventually canned.
Instead of finding some realistic product targets and then setting a budget around what resources will be needed to attain those goals, the budget is allocated based solely on desired revenue and somehow hiring all these people without a clear plan will make that happen.
The number I have heard (from maybe 15 years ago) is that "2/3 of corporate mergers meet their goals." That was across a range of industries. Tech in 2020 might be different, there is a certain kind of mindlessness around acquisitions that happens around technology sectors. For instance, AOL bought Time Warner, regretted it, AT&T bought Time Warner, now the vulture funds like Elliot Capital Management are circling and they will have the CEOs head.
AT&T's mistakes were made by people who were more technical than MBA, but it's a consistent trope that "Google bought Company A" and then Company B brings their salespeople to muster the next morning to inform them what's going to happen at Company A, and the next sales retreat is fistbumps all around.
I wonder if we could put a break on mergers and acquisitions, not for the sake of communities, workers, and competition, but just for the poor stockholders!
Again it depends. Look at Church & Dwight, more known as the Arm & Hammer company. Their strategy of expanding all over the grocery store cleaning isle seems to have worked as they have a whole team of brands and brands at different price premiums. They have bought smaller cpg companies successfully
yes, certainly there are successes, but it's good to realize that that's only about half of attempts. if you're an involved shareholder, it's eminently reasonable to be skeptical of executive strategies that feature m&a centrally.
And even with the half rate "successes" ( Success in terms that is actually defined by the executive themselves ), most of them were done with a miraculously talented team that normally has little to do with the executives.
Because in an ideal world you expect M&A to provide synergy, but it is often where 1+1 = 1.5 or less, and at best successes are 1+1 = 2. What people should be aiming for is 1+1 = 2.5 or more.
And recent Intel ( Post Andy Grove or Post Patrick Gelsinger ) has a track record of both poor strategic decision, acquisition and execution.
Note: That is M&A with respect to diversification.
Perhaps when the acquisition target is a smaller but fairly mature company who is being asked to scale existing processes. I think this is much more difficult for a startup when you’re essentially asking them to scale both their output and shrink their time to maturity by orders of magnitude.
But I didn't ask "when was there ever an acquisition of a small startup by a large incumbent that was successful?" I'm specifically thinking of when that small startup was then flooded with budget and people to meet aggressive growth targets.
I don't know about every item in your list, but I do know at least in some of those cases that the acquired company was largely left alone. They were perhaps given some integration goals, but the parent company specifically "left them alone" to not kill the secret sauce that made them successful in the first place.
From what I see in the B2B space, acquisitions are a major avenue for quick innovation. Large corporations buy smaller startup, do some integration into existing solutions and come out with a new product (If they don't just keep the new company as a separate costumer-facing identity (Look at NetApp + SolidFire, or the whole Dell-EMC conglomerate)
I'm so thankful that big corporations tend to be lumbering, clumsy giants like this - otherwise there wouldn't be room for any smaller companies, and we'd already be at the point where all the commerce in the world was owned by a handfull of megacorps (we might still get there).
>> So the message was "Get on and hire, because it won't be like this forever, and if you fail your hiring target, you'll be stuck delivering big projects with no staff".
Pay attention kids. There are times when a company is investing in the future. Do exactly that. Buy equipment that will be useful down the road because it won't last.
One place I worked had some awesome custom test stands for the product we were developing. I wondered how they ever came to be since money seemed tight and we actually had some layoffs. Development had been going on since years before I got there and they hadn't landed a customer yet. We rebooted the design for the 3rd time and did finally win. The business became very significant to the company. Sure, they let a small team keep trying but we never could have made it without that initial investment in equipment/infrastructure. Spend the money when it's available, but try to make it count.
The question is: What is the payoff for Intel from all these acquisitions? I've not see any that clearly were a net profit.
Mobileye: Paid $17B for it.
Altera: Paid $17B for it. At the time, Altera had a net profit of $0.5B. With something like $2B in assets, this will take 30 years to break even. How much growth are they expecting to offset that?
McAfee: Paid $8B for it, sold it for later. I'd love to see if they ever made back the amount they paid for it.
Nervana: They recently announced they're dropping the HW they produced. Time will tell if they can adapt the Nervana SW stack to the other ML company Intel just bought.
Infineon: Would love to see the numbers on whether they made a net profit on it. They may actually have with the sale to Apple.
WindRiver: Same question.
Yes, big companies should do some big bets. But bets are to make a net profit in the long run, with several failures along the way. With the amount Intel Capital has dumped in buying companies, and adding in the costs involved while holding them (employee salaries, capital expenses, etc), has Intel made money or lost money?
Thank You for writing pretty much all I want to write.
>Infineon: Would love to see the numbers on whether they made a net profit on it. They may actually have with the sale to Apple.
Likely not. Intel paid $1.4B for Infineon, invested billions into it every year, could not catch up to Qualcomm. Finally got the Apple contract by reducing price. And when they did it took out many of their 14nm capacity that could have been better used for something else. The $20 per pieces also did not really cover their R&D as stated in the court case against Qualcomm. In the end Apple got most of their R&D for $1B, and had to force to sign a license with Qualcomm or faced with the possibility of Intel shutting down their Modem business and they are left with no modem for iPhone.
It's a bet on every car on the planet having an autonomous ai in some foreseeable future.
Which equals to usually three pieces of redundant silicon with high availability failover, functional safety on each, high computer power needs and hard real time requirements (translating to further companion acceleration silicon).
Intel would love that to be Intel instead of some arm, riscv, Nvidia, whathaveyou...
The global vehicle production is in the order of 100mio vehicles per year, and drive ai is a growth market.
Don't forget Havok (sold to Microsoft, presumably at a loss), Recon (shut down), and Lemoptix and Composyt Light Lab (combined into Vaunt and shut down).
at first they thought it would sell quad-core processors, but discriminating consumers learned that third-party antivirus makes your machine slow, unstable, hot and runs down the battery. Thus it damaged Intel's brand.
> at first they thought it would sell quad-core processors, but discriminating consumers learned that third-party antivirus makes your machine slow, unstable, hot and runs down the battery. Thus it damaged Intel's brand.
I don't get the anti-virus market at all. Granted all AV products are probably horrible and I haven't used a third party AV on a personal machine ever since about 2009-ish [0] but back in college I distinctly remember the consumer AV stuff was so much taxing on performance compared to "enterprise" stuff. I don't think there was ever a technical reason for this, was there? So, why?
Now I do remember subversion checkouts or anything file intensive was pretty bad (and it still is if you don't have a decent SSD) but my impression is that the enterprise av was better than consumer? Am I misremembering this?
[0] I think I switched to Microsoft Security Essentials or some variant of it if I recall correctly.
That would be down to how they do accusitions, it's not just what they are paying, it's also backedup with a plan and investment. So you get a couple of years of nice big budget and the expectations would be that it would yeild back within that period. Now if those expectations play out, not just the product doing as they expected, but also the customers buying into it as expected. Those two are sepeate beasts. So either the company lives up to expectations and everything aligns and carry's on, or it gets a restructuring, which can take many forms and can happen more than once.
But like anything new, gets fat marketing budget, then after that grace period runs out, they are expected to be self sufficient and yielding a return.
Then even if everything goes well, those expectations and returns will also grow and eventually things change.
But initially, the honeymoon period is always fun, bit like first year at university, then things start to get real, real fast.
Completely agree. Had a friend start a career at Intel, moved to Apple because he saw it coming.
In case anyone isn't totally clear on why Intel is in such trouble: they made a lot of money selling high-end CPUs for desktops, and even higher-margin parts to datacenter operators.
On the PC front, when I was in high school (late 90s/early 2000s) it was common to have the computers on a "3 year replacement cycle" where they'd replace 1/3 or 1/4 of the computers every single year. Compare that to today where "2009 era" iMacs (!!) are floating around, ready for purchase at 200-300 dollars. There's no growth here.
On the mobile side, Intel had the chance to make chips for Apple but said no. They couldn't make the numbers work volume-wise. It's an interesting story, read it if you haven't.
Then comes datacenters. High-performance, high-power has been an Intel stronghold for 10-20 years but the whole industry is consolidating into AWS/Amazon/Azure who will probably make their own CPUs at some point, if not their entire own PCs. Either way this is bad for Intel: they face either someone who doesn't want their product, or a much larger buyer with much more purchasing leverage than a corporate IT department buying in small quantities. Same deal with storage -- the larger operators aren't paying the crazy EMC/NetApp margins those guys could cram down corporate IT's throats. Amazon has already announced plans to make ARM EC2 instances for reasons of both power and not paying Intel's crazy margins.
The thing Intel still does have is world-leading fab capabilities. I remember back in the 90s Intel always outperfomred AMD by running at much higher clockspeeds, and they could do it because of their superior process technology. But that allowed AMD to run laps around them in microarchitecture, which is showing with Ryzen. I don't see Intel becoming a contract manufacturer soon but I'm sure they've discussed it internally.
Lets look at least the Highend cellphone space you have Qcom and apple.Even Samsung doesn't really compete with Qcom. Qcom has a moat for their mobile business which is their radio patents. Even apple uses the Qcom radio chips, but they can because they have their insane margins.Lets look at the biggest semiconductor company in the world tsmc they generated about 4bil in Free cash flow. Intel generated about 18, and going after the low cost market is not a good idea. The best option would be for intel is just to return money back to the shareholders.
"Hey, we should start selling ARM CPUs. I know that the profit margin is two orders of magnitude less than x86 processors, but we'll sell an order of magnitude more chips."
The only question would be whether they would let that guy leave through the door under his own power.
The logical solution is to ride the gravy train until it tanks completely.
It's better than that. Intel was an ARM licensee for a brief period after they acquired StrongARM from DEC and developed Xscale. If they had any vision that could have been leveraged into a best in class low power mobile platform.
Intel can sell x86 with 50-80% margin + 80-90% market share.
Or sell arm compete with TW, China ARM licensee at 10-25% margin for 5-15% market share. And you don't have any competitive advantage like Qualcom with IP on CP.
>If you're Intel CEO, how are you going to choose?
Both. There is no reason why it is one or the other. And this is not just in hindsight, I said this even before BK becomes the CEO.
In the short term, or at leats in the terms of CEO thinking, it is the best to concentrate on the money making platform. In the longer term, this will come back and haunt them.
> Both. There is no reason why it is one or the other.
Oh, now that's just not true.
Intel fabs were at 90+% capacity (it's why they outsourced a lot of the support chips). You would have to bump an x86 in order to produce an ARM.
This is one of those Innovator's Dilemma moments. Even IF we concede that ARM will eat x86 (and I don't concede that without a lot more evidence), there is no moment where switching from making x86's makes sense until Intel is almost near death.
That is their fault for delay in the 10nm Fab not online in Israels, not building more capacity in existing Fabs, and not doing better Forecasting. All of these happened during BK's tenure.
Intel Custom Foundry was a little late, but that was 2012, they have achieved little to nothing during its 5 years of existence. The point is, if you do not have a product offering that is inside those 1.5B annual shipment of Tablet / Smartphone SoC, Modem Baseband and WiFi, at least offer to Fab those product line since Intel and x86 is not competing in that segment. Instead Samsung and TSMC got that market volume and is now able to sustain the leading edge R&D.
20 years ago they could have beat Qualcomm who was still just doing phone backends. They could have bought them too. Maximizing profit on all the things can leave exposed when you ignore emerging markets you could have been the dominant player on.
I’m just wondering why they have a billion to spend on this. Clearly they can walk and chew gum at the same time, yet being so under fire directionally makes me think they should be more focused on getting their core business right instead of chasing extremely expensive pipe dreams. Those side distractions almost always get cut at Intel...
Agreed, Intel seems to buy companies and then destroy or borg only small bits leaving everything else to rot.
But such purchases are good PR for the shareholder perception and almost gets down to churning purches thru the books on the surface, look good at that level of perception.
Agreed, They should focus upon what they have and with that, perish the thought they focused upon there FPGA with toolchains and support that could drive it into that and many markets.
If Intel added a small FPGA to their CPU's and they could - that could drive a whole new market of development and one in which they are placed to capitlise upon what IP they do have. Certainly an FPGA would make for a good mix with AI, if Intel would make such tools and open up FPGA's more to users and get that adoption going. Then the whole CPU/GPU race may well change into their favour.
Those acquisition that are in terms of actual silicon production are at least still running, because that is part of Intel's expertise. Those that has absolutely nothing to do with Silicon production are all being run to the ground.
They should have diversify into either into more silicon products or Custom Fabs, instead their Custom Foundry was a complete mess. What Intel dont realise was their competitor isn't actually AMD or Nvidia, but TSMC and Samsung. The market has no problem paying premium for leading edge node, leading edge processor design and software support. The problem is Intel lost on the first two and lost the third with respect to CUDA.
At least the new CEO seems to understand the problem better.
Wait... let's rewind. Why does Intel have an "Israeli automotive hub" from acquiring "Mobileye, the autonomous driving company" in 2017?
It makes sense when a large company acquires a smaller one, or founds a division from scratch, because it can provide that smaller company/company with uniquely valuable resources. E.g. Google/Alphabet founding Waymo because of Google's existing expertise in ML.
But how does expertise in processors translate to any kind of advantage for autonomous driving? Presumably the microprocessors in a self-driving car are one of the most commoditized parts... who cares if it's Intel, AMD or ARM?
Does anyone "in the know" have a clue around Intel's strategy here?
I can only guess they want to repeat the "Intel Inside" marketing campaign of the 1990's -- i.e. if they lost mobile, they don't want to lose cars too. But then wouldn't they want to form strong subsidized partnerships, the way they did with Intel Inside? Developing their own technology seems like it would make the autonomous car industry view them as competitors, not partners, and thus less likely to use their chips.
NVidia has product lines specifically for this field. Tesla proudly announced to the world that they made their own chip for autonomous driving. "Autonomous driving" research/development right now means lots of Machine Learning, image processing, ... High-performance stuff that benefits from GPU-like designs, all the while the systems for that need to be reasonably energy-efficient and build in a way that they'll survive in an automotive setting. They're not a competitor to car companies because car companies typically don't want to design chips themselves, they'll happily buy them if they can (Tesla is a really odd one out in that regard, and even they'll use superior hardware made by someone else if it makes sense). If it takes off across manufacturers, Intel very much would like a slice of that pie instead of leaving it all to NVidia&Co.
Sure, NVidia has chips that fit the needs of that field. Intel though has bought the actual company.
My only explanation is that Intel may believe they would have trouble breaking into that market with their own chips, and buying the company means they can make it dogfood Intel's products. If successful, it strengthens their sales argument to others.
Their competitor here is Nvidia, which has quite good hardware. Also, yes the ISA doesn't matter that much in the field, because no 3rd-party user-installed apps on automotive by design...
MobilEye's key advantage was they fabbed their own chips. Back when you couldn't really run a neural net in a car, they essentially kind of hard coded the evaluation of a certain neural net into a special chip.
Not sure I described this 100% accurately, but it's something like that. Hence the Intel acquisition. Their tech is pretty good honestly, used to power Tesla Autopilot 1 before they decided to build their own solution.
Actually one of Mobileye’s big strengths is their custom chips. They are purpose built and very efficient, which allows them to mount them under the windshield (important for their main business, ADAS), and they are built with the needs of the software in mind. I don’t know if Mobileye uses any Intel fabs, but it’s not such a crazy stretch to say one could help the other.
Intel does more than processors though. OpenCV itself originated at Intel and they still have an image processing library and several models of depth cameras with an SDK for face recognition, a physically based renderer and probably many other stuff.
Interesting acquisition by Intel, but also equally surprising valuation at $1B. I thought Moovit was entirely crowdsourced and made up primarily of volunteers[1]... maybe that changed recently?
Also the article claims “800 million riders” which is astounding for an app that is currently #36 in the navigation category (iOS App Store). Unless this app is more popular than TikTok everywhere outside of the US they’ve gotta be playing fast and loose with those usage statistics.
Moovit is for public transportation and that is pretty much why it isn't big in the US. Here it is #3 in navigation after Google Maps and Waze.
I couldn't live without it anymore, the city has probably around 500 bus lines. Making 3 combinations isn't unusual and it works to the point where I don't worry about researching a new route. Whenever you aren't commuting it is your best option to finding a route.
Sometimes I spend a weekend in smaller cities I don't know well. They have bus lines which you can't find online anywhere. But everything is on Moovit. This has saved me so many hours, I feel like I should be paying for it.
I nominate Spectron. It died twice I think: the corpse was sold to Dialogic, then Dialogic was bought by Intel later on, and then there was more spinoff/merger pinball.
The original idea was for Intel to wrest control of multimedia from Microsoft and create an Intel controlled multimedia subsystem. They didn't actually acquire Dragon for the speech recognition part, but Dragon's fate was even worse.
This appears to be a very desperate move by a company just realizing the desperate situation they find themselves in. I agree with others sentiment that they need to find a way to get back to doing well what it is they know, or used to know.
Yeah Intel really has their back against that wall, after all they are one of the most highly capitalized businesses in human history, just recorded their largest annual revenue in the history of the firm, and earn a profit of only about $80000 per minute. Pure desperation down there in Santa Clara.
Financial results are always a lagging indicator in tech. That doesn't mean they aren't important, just that you can coast on an existing product line for a really long time (decade+) and then everyone's "surprised" when unit sales decline, and there's nothing else in the pipeline.
An example of this: Sears. Dominant retailer, completely missed the boat on ecommerce, now bankrupt. Literally.
Sears missed the boat on ecomerce, but that isn't why they are gone. Walmart missed the boat too at first, but they are still big because their manageme was good. They also have adjusted overtime such that now my family does most of our shopping there (once this pandemic is done I'm looking forward to going back to the local stores I used to go to.)
Let me grab this occassion to draw attention to a lesser known feature of Citymapper (not everywhere alas): it can make a public transit plan augmented by a cab. This is something I have been doing without an app myself at my home city in a very clear demonstration of the "time is money" principle. This is one of the true differentiating capabilities between Citymapper, Moovit and Google Maps, otherwise they would be quite similar.
I actually hate that feature and can’t figure out how to turn it off. It thinks getting me within 1-2 miles of my destination and hailing a car is better than being on a train for 20% longer but being much closer to where I actually want to go. I outright stopped using the app because it’s too annoying to make sure the suggested route doesn’t involving getting off a train and catching a car.
The other day I was defending buybacks and dividends on HN as policies that protect owners from managers. These kinds of acquisitions are exactly what I had in mind.
I’ve used Transit and Google Maps and to me Transit’s suggestions are closer to how a human considers transit options.
Google Maps seem strangely over optimized for things like minimum walking at the beginning of the route. Plus the Google Maps “how far away the bus is” doesn’t update based on new data (from the city bus trackers or crowdsourced data) unless you close the app and re open
They do solve the same problem but Google Maps public transportation data is not crowdsourced, if there is no GTFS feed then it has nothing. Living in a 3rd world country where we have many buses but little data I can tell you that Moovit works much much better here.
This in my opinion is just a bailout acquisition from Intel to Moovit, made to make Intel look more innovative in the face of a stagnating processor division while bailing out Moovit from bankruptcy.
Intel screwed up and missed the mobile phone revolution.
They’re probably trying to make sure they don’t miss out on the autonomous car revolution. Which will likely require more chips, in both vehicles and infrastructure.
But this is a huge gamble, as an infrastructure upgrade like this will cost billions or trillions. And it may likely never happen. But if and when it does, then Intel may be standing on the next wave of huge spending for its technology.
These acquisitions almost always end up being run into the ground because Intel is so used to turning the crank on their mature (read: awesomely ubiquitous software support) CPU architecture that they don’t want to spend much on R&D on things that aren’t proven money makers. But they keep trying to div diversify in a very half-assed way because they have such a high market share in CPUs (not that it is without legitimate threat from AMD) that to continue to grow they want to hit new markets.